cur8-skills
v0.1.0
Published
CMO judgment for your AI agent — Cur8's marketing skills for Claude Code, Cursor, and Codex: strategic diagnosis, positioning, outreach, investor updates, press releases, campaign plans.
Maintainers
Readme
Cur8 Skills
CMO judgment for your AI agent. Cur8's marketing skills for Claude Code, Cursor, and Codex — the working standards of the Cur8 AI CMO, installable in one command.
npx cur8-skills installThen, inside your agent:
/cur8 diagnose my company
/cur8-outreach a 3-email sequence for plant heads at steel producers
/cur8-investor-update from these bullets: ...The skills
| Skill | What it does |
|---|---|
| /cur8 | Router — ask for any marketing job in plain English |
| /cur8-diagnose | The Atlas strategic diagnosis: research-first, 9 lenses, a CEO Brief you can read in 60 seconds |
| /cur8-positioning-gap | Homepage/pitch teardown: the story you tell vs the one that creates demand |
| /cur8-icp | Prose → precise ICP with a ready-to-run prospecting filter set |
| /cur8-outreach | Cold email & sequences under Cur8's human-voice rules (no AI tells, 4-sentence cap, grounded claims) |
| /cur8-investor-update | Investor updates with metrics discipline and honest lowlights |
| /cur8-press-release | Wire-structure releases: sourced claims only, FLS hygiene |
| /cur8-competitor-teardown | Steelman first, then the flank |
| /cur8-campaign-plan | Goal → channels → budget bands → kill criteria |
The two rules every skill enforces
- Grounding — specific claims (customers, numbers, outcomes) come only from material you provide. With nothing to cite, the skills write around the gap. They never invent a proof point.
- Human voice — no em-dashes, no AI vocabulary, short sentences, outcomes over adjectives.
These are the same standards the Cur8 platform enforces in production — where drafts become sent, tracked, reply-aware, compliance-gated campaigns. Skills draft anywhere. Shipping happens through the gate → cur8.dev
License
MIT. Read them, fork them, make them yours.
